{"id":"arxiv/2609.07077","url":"https://arxiv.org/abs/2609.07077","tags":["math.ST"],"score":0,"title":"Network recovery from aggregated relational data: An impossibility theorem","domain":"math.ST","source":"arxiv","authors":["Yen-hsuan Tseng"],"reviews":0,"subject":"math.ST","summary":"We prove an impossibility theorem for uniformly consistent recovery of labeled edge probabilities from complete aggregated relational data, even when the population count law identifies every probability. For any known partition into two equal trait groups, we consider an independent-edge logistic network with a balanced rank-one signal and unknown activity effects with bounded total dyadic energy. The signal amplitude is known and fixed at a small positive value. Every node reports its counts to both groups. The minimax mean squared error for the probability matrix remains bounded away from…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-09-07","reactions":{}}